Boosting Higgs Discovery
Graham Kribs
University of Oregon
Santa Fe 5 July 2010
Punchline
Mass (GeV)0 20 40 60 80 100 120 140 160 180 200
-1Ev
ents
/ 8G
eV /
30fb
0
20
40
60
80
100
120
140
Mass (GeV)0 20 40 60 80 100 120 140 160 180 200
-1Ev
ents
/ 8G
eV /
30fb
0
20
40
60
80
100
120
140 qqV+jetsVVV+Higgs
= 4.5BS/in 112-128GeV
(d)
BDRS developed a qualitatively new method,using “jet substructure”, to demonstratediscovery of h -> b,bbar at the LHC for
light Higgs masses with S ! 4 for 30 fb-1
We find that light MSSM Higgses can also be discovered using BDRS technique
(but with a few additional tricks of our own) applied to the sparticle production and
decay into h (& H,A).
Could be first h discovery mode beforeh -> !!, h -> ""!!
“We”Adam
Tuhin
Michael
Fermilab postdoc
UO->UW postdoc
UO postdoc
“Discovering the Higgs Boson in New Physics Events using Jet Substructure”0912.4731 [PRD]
“Discovering Higgs Bosons of the MSSM using Jet Substructure”1006.1656
Martin
Roy
Spannowsky
Outline
• Higgs in SM
• Boosting & Jet Substructure
• Higgs in SUSY
• Boosted SUSY Higgs
• Summary
Outline
• Higgs in SM
• Boosting & Jet Substructure
• Higgs in SUSY
• Boosted SUSY Higgs
• Summary
Branching Ratios of SM Higgs
HDECAY
Branching Ratios of SM Higgs
Between115 < mh < 135h -> b,bbardominates
!!
(GeV)Hm100 120 140 160 180 200 220
expe
cted
sig
nific
ance
0
2
4
6
8
10
12
14
16
18Combined
4l→ (*)
ZZγ γτ τ
νµν e→WW0j νµν e→WW2j
ATLAS-1L = 10 fb
ATLAS TDR 2009 - 10 fb-1
100 120
0
2
4
6
8
ATLAS TDR 2009 - 10 fb-1
Combined 4l→
(*)ZZ
γ γτ τ
νµν e→WW0j νµν e→WW2j
CMS TDR 2006 - 30 fb-1
2,GeV/cHM
Sig
nific
ance
1
10
100 200 300 400 500 600
cuts!!"H opt!!"H
4l"ZZ"H#2l2"WW"H
jj#l"WW"qqH, Hl+jet"$$"qqH, H
!!"qqH, H
-1CMS, 30 fb
CMS TDR 2006 - 30 fb-1
1
10
100 200
C
cuts!!"H opt!!"H
4l"ZZ"H#2l2"WW"H
jj#l"WW"qqH, Hl+jet"$$"qqH, H
!!"qqH, H
Viable Channels for 115 < mh < 125
h -> !!
h -> ""
BR(h -> !!) ! 1-2 x 10-3
BR(h -> "") ! 5-7 x 10-2
Until 2008, was thought that the largest one,BR(h -> b,bbar) ! 0.7-0.9
was lost in the QCD background.
Outline
• Higgs in SM
• Boosting & Jet Substructure
• Higgs in SUSY
• Boosted SUSY Higgs
• Summary
Common Lore
p p
qbar
q
(for this theorist)
Common Lore
p p
q
“one hard parton per jet”qbar
Parton to Detector...
q
More complicated...
q
Parton shower; hadronization/decay...
More complicated...
q
Parton energy -> Jet energy
Need to capture radiation
Even more complicated...
q
Jet cone contains non-parton energy(underlying event)
Boosted Jets
p p
q
e.g. pp -> Zhqbar
Z -> l+l-
h -> q,qbar
Boosted Jets
q
Jet cone has energy from twohard partons and their showers!
qbar
Boosted Jets
q
Enlarge jet cone to capture hard partons and shower radiation
qbar
Boosted Jets
q
Boosted Advantage: Objects with high pT have excellent jet energy resolution:
qbar
(ATLAS TDR,cone jets.)
(δE
E
)
jets
∼=0.6√
E/GeV+ 0.03
Boosted Jets
q
Large Cone Disadvantage: Underlying event significantly contaminates “fat jet” energy
qbar
BDRS(Butterworth, Davison, Rubin, Salam; 0802.2470)
Proposed a technique that involved “jet substructure” and “filtering”
to discover
h -> b,bbar
for 115 < mh < 125 GeV range.
BDRS Basic Premise:Lose cross section by requiring Higgs is boosted
pT(h) > 200 GeV(only 5% of Zh/Wh production at 14 TeV LHC has this level of boost)
Gain on signal/background through therelative uniqueness and characteristicsof the “fat jet” over QCD background
BDRS Requirements
• trigger! Zh or Wh with leptonic Z/W decay• pT(h) > 200 GeV• several (standard) cuts to help reduce background
• cluster jets with inclusive Cambridge/Aachen R = 1.2• “fat jet” is b-tagged• “mass drop” from jets to subjets• “symmetric” pT from jets to subjets• two b-tags of two highest pT subjets• filter the subjets
Jet Decomposition
b-tagged jet
with jet mass: mj
Jet Decomposition
undo clustering
j1
j2
In C/A, this came from two jets with masses mj1 , mj2
repeat unclustering:
j21
j12
j22
Jet Decomposition
four jets with masses: mj11 mj12 mj21 mj22
j11
1) check for mass drop
j1
j2
mj1 < µ mj?0<µ<1 is a parameter
2) check “asymmetry”
y =min
“p2
tj1,p2
tj2
”
m2j
∆R2j1,j2
> ycut
d ( )2 d l R fycut is a parameter
Higgs Jet Identification
j1
j2
mj1 < µ mj?
BDRS used µ = 0.68
BDRS used ycut = (0.3)2
Higgs Jet Identification
y =min
“p2
tj1,p2
tj2
”
m2j
∆R2j1,j2
> ycut
d ( )2 d l R f
Expect drop forheavy particle decay
Tends to reject soft/colinear
QCDcontamination
1) check for mass drop
2) check “asymmetry”
3) are both subjets b-tagged?
j1
j2
yes - continueno - scrap this “fat jet”
Higgs Jet Identification
4) Filter the subjets to remove more UE:
• take 3 highest pT subjets (“third” captures leading parton shower gluon)• subjets Rj1,j2,j3 = min(Rbb/2,0.3)
Higgs Jet Identification
Rbb
Higgs Candidate Mass formed from 3 highest pT subjets
j21j12
j22j11
Mass (GeV)0 20 40 60 80 100 120 140 160 180 200
-1Ev
ents
/ 8G
eV /
30fb
0
20
40
60
80
100
120
140
Mass (GeV)0 20 40 60 80 100 120 140 160 180 200
-1Ev
ents
/ 8G
eV /
30fb
0
20
40
60
80
100
120
140 qqV+jetsVVV+Higgs
= 4.5BS/in 112-128GeV
(d)
BDRS 0802.2470
BDRS Result
• LHC 14 TeV; 30 fb-1
• HERWIG/JIMMY cross-checked with PYTHIA with “ATLAS tune”• 60% b-tag; 2% mistag• no smearing
ATL-PHYS-PUB-2009-088 (Aug 2009)
]2Higgs mass [GeV/c0 20 40 60 80 100 120 140 160 180 200
!1Ev
ents
/ 8G
eV /
30fb
02468
101214161820
]2Higgs mass [GeV/c0 20 40 60 80 100 120 140 160 180 200
!1Ev
ents
/ 8G
eV /
30fb
02468
101214161820
HiggsV+jetstopVV
Total S = 13.5 B = 20.3Range 112!136GeV
ATLAS preliminary(simulation)
(a)
]2Higgs mass [GeV/c0 20 40 60 80 100 120 140 160 180 200-1
Even
ts /
8GeV
/ 30
fb0
2
4
6
8
10
12
14
16
]2Higgs mass [GeV/c0 20 40 60 80 100 120 140 160 180 200-1
Even
ts /
8GeV
/ 30
fb0
2
4
6
8
10
12
14
16ZjZZHiggs
Total S = 5.3 B = 12.2Range 104-136GeV
ATLAS preliminary(simulation)
(b)
ATLAS Simulation
Outline
• Higgs in SM
• Boosting & Jet Substructure
• Higgs in SUSY
• Boosted SUSY Higgs
• Summary
h is light
Carena, Haber Higgs report 2002
Supersymmetrize Higgs Interactions
H H H H~ ~ ~
B,W+-0 ~ ~supersymmetrize
g’,g g’,g
B,W+-0
Gauginos can decay to Higgs I
“Goldstone region”When mA large,! 25% of time
decay to h ! 75% of time to longitudinal W/Z
~ ~B,W+-0 H+-0~
h,z,w+-
(if kinematically open)
Kinematical requirementmW > µ + mh
> 100 + 120 = 220 GeV
Gauginos can decay to Higgs II
H~
h
(Higgsino NLSP; Gravitino LSP)
Kinematical requirementµ > mG + mh
> 120 GeV
G~
~
Squark Production to Gauginos
~ ~B,W+-0 H~
hq
q
~
typical #(squarks)14 TeV ! several pb!!
(which then decay to Higgs)
MSSM Higgs• Higgs is light; h -> b,bbar always present!
• New production sources of Higgs from superpartner cascades can have large cross sections and large decay rates
• Heavy superpartners decaying to light Higgs implies boost is often large (! 50%!)
Ideal for Boosted Higgs Analysis!
Outline
• Higgs in SM
• Boosting & Jet Substructure
• Higgs in SUSY
• Boosted SUSY Higgs
• Summary
�2 �1 0 1 20
10
20
30
40
50
60
70
M1�Μ
MSSM (Neutralino LSP)
pT(h) > 200 GeV
pT(h) > 300 GeV
Boosted Fraction
msq = 1 TeVM2=2M1
Problem with SUSY production:Supersymmetric events tend to be
“busy” with a lot of hadronic activity
from squarks/gluinosand associated parton showers.
Can have extrahard subjets in “fat jet” cone!
~ ~B,W+-0 H~
hq
q
~
Similar problem also with t,tbar,h! Plehn, Salam, Spannowsky
Our New Step: Maximize “Similarity”1) Do BDRS procedure.
Si =min
(p2
tj1, p2
tj2
)
(ptj1
+ ptj2
)2 ∆Rj1j2
3) Choose 3 highest pT jets (2 highest b-tagged) from stage which maximizes “S”
2) At each stage of unclustering, calculate “S”:
This helps improve the efficiencyof finding the Higgs
j11 j21
j12
j22
Signal: SUSPECT2 PYTHIA6.4
Background: ALPGEN PYTHIA6.4 underlying event:ATLAS tune
• All final-state hadrons grouped into cells of size (∆η × ∆φ) = (0.1 × 0.1)
• Each cell is rescaled to be masslessthis models detector response (Thaler, Wang ’08)
b-tagging:
jet-photon fake rate:
60% efficiency,
.1%
2% fake rate
jet gymnastics performed using FastJet (hep-ph/0512210)
Simulations details...
MSSM (Neutralino LSP) Point 1A
/ET > 300 GeV, HT > 1.0 TeV
|µ|M1
M2
750 GeV
600 GeV
300 GeV−250 GeV
mQ̃3
1 TeVmQ̃1,2
150
+ lepton veto
Improvement with Similarity
MSSM Point 6
Can even discover heavier A,H states!
|µ|M1
M2
750 GeV
600 GeV
300 GeV−250 GeV
mQ̃3
1 TeVmQ̃1,2
150
now, mA = 200 GeV
MSSM (light bino) Point 3
(Relic density of LSP gives right thermal abundance)
|µ|M1
M2
750 GeV
600 GeV
300 GeV−250 GeV
mQ̃3
1 TeVmQ̃1,2
150 GeV
400 GeV
200 GeV|!|
Summary• Jet substructure techniques can find boosted Higgs h -> b,bbar, even in busy environment
• MSSM h automatically light, could have large rate from squark production and cascade decay
• Cascade decay implies large fraction boosted!
• Could discover h (and/or A,H) well before conventional modes -- less than 10 fb-1!!